AutoML Leaderboard

Best model	name	model_type	metric_type	metric_value	train_time	single_prediction_time
	1_Linear	Linear	logloss	1.66983	47.02	0.2917
	2_Default_LightGBM	LightGBM	logloss	0.62357	5.63	0.2773
	3_Default_Xgboost	Xgboost	logloss	0.611399	421.02	0.285
	4_Default_CatBoost	CatBoost	logloss	0.616243	55.2	0.3436
	5_Default_NeuralNetwork	Neural Network	logloss	2.23502	8.64	0.3918
	6_Default_RandomForest	Random Forest	logloss	0.6278	42.73	0.3675
	11_LightGBM	LightGBM	logloss	0.634161	5.48	0.2782
	7_Xgboost	Xgboost	logloss	0.625457	406.41	0.2809
	15_CatBoost	CatBoost	logloss	0.617327	99.51	0.3408
	19_RandomForest	Random Forest	logloss	0.616654	27.14	0.3325
	23_NeuralNetwork	Neural Network	logloss	2.28946	9.04	0.3891
	12_LightGBM	LightGBM	logloss	0.663489	5.66	0.2851
	8_Xgboost	Xgboost	logloss	0.630069	409.3	0.2822
	16_CatBoost	CatBoost	logloss	0.629594	99.23	0.3626
	20_RandomForest	Random Forest	logloss	0.615712	22.55	0.3197
	24_NeuralNetwork	Neural Network	logloss	0.893568	9.77	0.3886
	13_LightGBM	LightGBM	logloss	0.635894	6.07	0.2787
	9_Xgboost	Xgboost	logloss	0.631836	404.38	0.2839
	17_CatBoost	CatBoost	logloss	0.606855	67.95	0.348
	21_RandomForest	Random Forest	logloss	0.622754	25.19	0.3309
	25_NeuralNetwork	Neural Network	logloss	1.80514	10.13	0.3891
	14_LightGBM	LightGBM	logloss	0.631686	7.14	0.2783
	10_Xgboost	Xgboost	logloss	0.693147	408.67	0.2915
	17_CatBoost_RandomFeature	CatBoost	logloss	0.640051	61.36	0.3448
	17_CatBoost_SelectedFeatures	CatBoost	logloss	0.586664	3.38	0.0198
	20_RandomForest_SelectedFeatures	Random Forest	logloss	0.594476	4.78	0.0399
	2_Default_LightGBM_SelectedFeatures	LightGBM	logloss	0.615567	3.19	0.0174
	24_NeuralNetwork_SelectedFeatures	Neural Network	logloss	0.799705	3.77	0.0271
	26_CatBoost_SelectedFeatures	CatBoost	logloss	0.591951	3.51	0.0202
	27_CatBoost_SelectedFeatures	CatBoost	logloss	0.60816	3.79	0.0201
	28_RandomForest_SelectedFeatures	Random Forest	logloss	0.592273	5	0.0392
	29_CatBoost	CatBoost	logloss	0.618606	52.27	0.3502
	30_Xgboost	Xgboost	logloss	0.625119	405.85	0.2858
	31_LightGBM_SelectedFeatures	LightGBM	logloss	0.607325	3.62	0.0196
	32_LightGBM_SelectedFeatures	LightGBM	logloss	0.610673	3.64	0.0173
	33_RandomForest	Random Forest	logloss	0.618311	23.64	0.3236
	34_LightGBM	LightGBM	logloss	0.630705	7.85	0.2778
	35_LightGBM	LightGBM	logloss	0.631686	7.95	0.2812
	36_Xgboost	Xgboost	logloss	0.622813	404.59	0.2839
the best	Ensemble	Ensemble	logloss	0.576474	0.88	0.4025

AutoML Performance

AutoML Performance Boxplot

Features Importance

Spearman Correlation of Models

Summary of 29_CatBoost

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 6
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

51.6 seconds

Metric details

	score	threshold
logloss	0.618606	nan
auc	0.686672	nan
f1	0.6313	0.349467
accuracy	0.665698	0.510485
precision	0.857143	0.693178
recall	1	0.0502593
mcc	0.290176	0.586134

Confusion matrix (at threshold=0.510485)

	Predicted as 0	Predicted as 1
Labeled as 0	171	30
Labeled as 1	85	58

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of Ensemble

<< Go back

Ensemble structure

Model	Weight
17_CatBoost	1
17_CatBoost_SelectedFeatures	1
28_RandomForest_SelectedFeatures	1

Metric details

	score	threshold
logloss	0.576474	nan
auc	0.762133	nan
f1	0.666667	0.432499
accuracy	0.732558	0.48995
precision	0.866667	0.71695
recall	1	0.0695731
mcc	0.44195	0.473934

Confusion matrix (at threshold=0.48995)

	Predicted as 0	Predicted as 1
Labeled as 0	170	31
Labeled as 1	61	82

Learning curves

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 20_RandomForest

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.7
• min_samples_split: 30
• max_depth: 7
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

21.9 seconds

Metric details

	score	threshold
logloss	0.615712	nan
auc	0.706328	nan
f1	0.652742	0.360186
accuracy	0.665698	0.431607
precision	1	0.720123
recall	1	0.0823607
mcc	0.344537	0.431607

Confusion matrix (at threshold=0.431607)

	Predicted as 0	Predicted as 1
Labeled as 0	125	76
Labeled as 1	39	104

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 9_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.1
• max_depth: 7
• min_child_weight: 25
• subsample: 0.9
• colsample_bytree: 0.6
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

403.7 seconds

Metric details

	score	threshold
logloss	0.631836	nan
auc	0.666458	nan
f1	0.63388	0.371623
accuracy	0.642442	0.548324
precision	0.787879	0.593492
recall	1	0.12641
mcc	0.287801	0.371623

Confusion matrix (at threshold=0.548324)

	Predicted as 0	Predicted as 1
Labeled as 0	183	18
Labeled as 1	105	38

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 20_RandomForest_SelectedFeatures

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.7
• min_samples_split: 30
• max_depth: 7
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

4.2 seconds

Metric details

	score	threshold
logloss	0.594476	nan
auc	0.733118	nan
f1	0.652632	0.489698
accuracy	0.712209	0.505232
precision	0.8125	0.777897
recall	1	0.0263305
mcc	0.406986	0.489698

Confusion matrix (at threshold=0.505232)

	Predicted as 0	Predicted as 1
Labeled as 0	154	47
Labeled as 1	52	91

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 5_Default_NeuralNetwork

<< Go back

Neural Network

• n_jobs: -1
• dense_1_size: 32
• dense_2_size: 16
• learning_rate: 0.05
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

8.0 seconds

Metric details

	score	threshold
logloss	2.23502	nan
auc	0.624743	nan
f1	0.602247	6.67617e-05
accuracy	0.633721	0.49831
precision	0.621622	0.999983
recall	1	1.75338e-10
mcc	0.250704	0.49831

Confusion matrix (at threshold=0.49831)

	Predicted as 0	Predicted as 1
Labeled as 0	135	66
Labeled as 1	60	83

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 7_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.075
• max_depth: 8
• min_child_weight: 5
• subsample: 1.0
• colsample_bytree: 1.0
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

405.8 seconds

Metric details

	score	threshold
logloss	0.625457	nan
auc	0.683993	nan
f1	0.614583	0.336257
accuracy	0.68314	0.519306
precision	0.784314	0.608961
recall	1	0.0893961
mcc	0.330634	0.530884

Confusion matrix (at threshold=0.519306)

	Predicted as 0	Predicted as 1
Labeled as 0	170	31
Labeled as 1	78	65

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 2_Default_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 0.9
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

5.0 seconds

Metric details

	score	threshold
logloss	0.62357	nan
auc	0.686637	nan
f1	0.637532	0.299511
accuracy	0.662791	0.4395
precision	0.761905	0.679846
recall	1	0.0808415
mcc	0.299321	0.4395

Confusion matrix (at threshold=0.4395)

	Predicted as 0	Predicted as 1
Labeled as 0	148	53
Labeled as 1	63	80

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 21_RandomForest

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.8
• min_samples_split: 40
• max_depth: 3
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

24.5 seconds

Metric details

	score	threshold
logloss	0.622754	nan
auc	0.703684	nan
f1	0.647887	0.370354
accuracy	0.668605	0.510011
precision	0.767442	0.556501
recall	1	0.1394
mcc	0.331172	0.376468

Confusion matrix (at threshold=0.510011)

	Predicted as 0	Predicted as 1
Labeled as 0	174	27
Labeled as 1	87	56

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 26_CatBoost_SelectedFeatures

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 6
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

2.9 seconds

Metric details

	score	threshold
logloss	0.591951	nan
auc	0.735205	nan
f1	0.674419	0.371462
accuracy	0.686047	0.480866
precision	1	0.784608
recall	1	0.0383728
mcc	0.388303	0.371462

Confusion matrix (at threshold=0.480866)

	Predicted as 0	Predicted as 1
Labeled as 0	155	46
Labeled as 1	62	81

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 24_NeuralNetwork

<< Go back

Neural Network

• n_jobs: -1
• dense_1_size: 64
• dense_2_size: 16
• learning_rate: 0.01
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

9.1 seconds

Metric details

	score	threshold
logloss	0.893568	nan
auc	0.652507	nan
f1	0.597403	0.0241688
accuracy	0.659884	0.510971
precision	0.66	0.869753
recall	1	2.34671e-05
mcc	0.288123	0.510971

Confusion matrix (at threshold=0.510971)

	Predicted as 0	Predicted as 1
Labeled as 0	152	49
Labeled as 1	68	75

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 31_LightGBM_SelectedFeatures

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 0.8
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

3.1 seconds

Metric details

	score	threshold
logloss	0.607325	nan
auc	0.709286	nan
f1	0.631313	0.290697
accuracy	0.686047	0.488579
precision	0.9	0.828996
recall	1	0.0320422
mcc	0.340025	0.488579

Confusion matrix (at threshold=0.488579)

	Predicted as 0	Predicted as 1
Labeled as 0	161	40
Labeled as 1	68	75

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 4_Default_CatBoost

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 6
• rsm: 1
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

54.4 seconds

Metric details

	score	threshold
logloss	0.616243	nan
auc	0.687959	nan
f1	0.623037	0.274348
accuracy	0.671512	0.567465
precision	1	0.752709
recall	1	0.020766
mcc	0.326109	0.654552

Confusion matrix (at threshold=0.567465)

	Predicted as 0	Predicted as 1
Labeled as 0	185	16
Labeled as 1	97	46

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 3_Default_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.075
• max_depth: 6
• min_child_weight: 1
• subsample: 1.0
• colsample_bytree: 1.0
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

420.4 seconds

Metric details

	score	threshold
logloss	0.611399	nan
auc	0.708712	nan
f1	0.63662	0.36052
accuracy	0.69186	0.495184
precision	0.916667	0.767444
recall	1	0.0475052
mcc	0.355286	0.495184

Confusion matrix (at threshold=0.495184)

	Predicted as 0	Predicted as 1
Labeled as 0	158	43
Labeled as 1	63	80

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 11_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 15
• learning_rate: 0.05
• feature_fraction: 0.8
• bagging_fraction: 0.5
• min_data_in_leaf: 50
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

4.8 seconds

Metric details

	score	threshold
logloss	0.634161	nan
auc	0.662909	nan
f1	0.609272	0.232769
accuracy	0.65407	0.458332
precision	0.85	0.635467
recall	1	0.054591
mcc	0.270795	0.458332

Confusion matrix (at threshold=0.458332)

	Predicted as 0	Predicted as 1
Labeled as 0	156	45
Labeled as 1	74	69

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 35_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 1.0
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

7.3 seconds

Metric details

	score	threshold
logloss	0.631686	nan
auc	0.659813	nan
f1	0.617169	0.24762
accuracy	0.642442	0.541502
precision	0.8	0.693257
recall	1	0.0456055
mcc	0.234163	0.345597

Confusion matrix (at threshold=0.541502)

	Predicted as 0	Predicted as 1
Labeled as 0	181	20
Labeled as 1	103	40

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 1_Linear

<< Go back

Logistic Regression (Linear)

• n_jobs: -1
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

46.4 seconds

Metric details

	score	threshold
logloss	1.66983	nan
auc	0.608009	nan
f1	0.601336	0.0010097
accuracy	0.616279	0.990348
precision	1	0.999988
recall	1	5.53986e-06
mcc	0.182939	0.999988

Confusion matrix (at threshold=0.990348)

	Predicted as 0	Predicted as 1
Labeled as 0	177	24
Labeled as 1	108	35

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 33_RandomForest

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.7
• min_samples_split: 30
• max_depth: 6
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

23.0 seconds

Metric details

	score	threshold
logloss	0.618311	nan
auc	0.703441	nan
f1	0.643243	0.37929
accuracy	0.662791	0.497338
precision	1	0.710595
recall	1	0.0824139
mcc	0.306738	0.37929

Confusion matrix (at threshold=0.497338)

	Predicted as 0	Predicted as 1
Labeled as 0	160	41
Labeled as 1	75	68

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 24_NeuralNetwork_SelectedFeatures

<< Go back

Neural Network

• n_jobs: -1
• dense_1_size: 64
• dense_2_size: 16
• learning_rate: 0.01
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

3.2 seconds

Metric details

	score	threshold
logloss	0.799705	nan
auc	0.66197	nan
f1	0.640449	0.294378
accuracy	0.651163	0.447762
precision	0.6	0.998235
recall	1	1.17194e-05
mcc	0.309217	0.294378

Confusion matrix (at threshold=0.447762)

	Predicted as 0	Predicted as 1
Labeled as 0	135	66
Labeled as 1	54	89

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 13_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 0.9
• bagging_fraction: 1.0
• min_data_in_leaf: 20
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

5.4 seconds

Metric details

	score	threshold
logloss	0.635894	nan
auc	0.663431	nan
f1	0.602532	0.344696
accuracy	0.665698	0.451738
precision	0.75	0.672436
recall	1	0.0999826
mcc	0.293646	0.451738

Confusion matrix (at threshold=0.451738)

	Predicted as 0	Predicted as 1
Labeled as 0	161	40
Labeled as 1	75	68

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 36_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.075
• max_depth: 8
• min_child_weight: 1
• subsample: 1.0
• colsample_bytree: 1.0
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

404.0 seconds

Metric details

	score	threshold
logloss	0.622813	nan
auc	0.689124	nan
f1	0.622785	0.344044
accuracy	0.686047	0.547778
precision	1	0.726162
recall	1	0.12656
mcc	0.336758	0.547778

Confusion matrix (at threshold=0.547778)

	Predicted as 0	Predicted as 1
Labeled as 0	174	27
Labeled as 1	81	62

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 2_Default_LightGBM_SelectedFeatures

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 0.9
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

2.6 seconds

Metric details

	score	threshold
logloss	0.615567	nan
auc	0.710643	nan
f1	0.636364	0.290416
accuracy	0.68314	0.505412
precision	0.703704	0.602933
recall	1	0.0344308
mcc	0.330121	0.505412

Confusion matrix (at threshold=0.505412)

	Predicted as 0	Predicted as 1
Labeled as 0	169	32
Labeled as 1	77	66

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 17_CatBoost_SelectedFeatures

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 7
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

2.8 seconds

Metric details

	score	threshold
logloss	0.586664	nan
auc	0.746791	nan
f1	0.653295	0.382873
accuracy	0.723837	0.489132
precision	0.8	0.709318
recall	1	0.0384375
mcc	0.421801	0.489132

Confusion matrix (at threshold=0.489132)

	Predicted as 0	Predicted as 1
Labeled as 0	166	35
Labeled as 1	60	83

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 30_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.075
• max_depth: 6
• min_child_weight: 5
• subsample: 1.0
• colsample_bytree: 1.0
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

405.2 seconds

Metric details

	score	threshold
logloss	0.625119	nan
auc	0.687089	nan
f1	0.626866	0.396066
accuracy	0.69186	0.543294
precision	0.84	0.693728
recall	1	0.123579
mcc	0.351499	0.547384

Confusion matrix (at threshold=0.543294)

	Predicted as 0	Predicted as 1
Labeled as 0	177	24
Labeled as 1	82	61

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 12_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.2
• feature_fraction: 0.5
• bagging_fraction: 1.0
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

5.0 seconds

Metric details

	score	threshold
logloss	0.663489	nan
auc	0.604408	nan
f1	0.598639	0.25607
accuracy	0.616279	0.621098
precision	0.703704	0.649582
recall	1	0.0440368
mcc	0.187413	0.373748

Confusion matrix (at threshold=0.621098)

	Predicted as 0	Predicted as 1
Labeled as 0	191	10
Labeled as 1	122	21

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 28_RandomForest_SelectedFeatures

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.7
• min_samples_split: 30
• max_depth: 6
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

4.4 seconds

Metric details

	score	threshold
logloss	0.592273	nan
auc	0.736023	nan
f1	0.644444	0.510373
accuracy	0.723837	0.519674
precision	0.8125	0.712595
recall	1	0.0128571
mcc	0.422706	0.519674

Confusion matrix (at threshold=0.519674)

	Predicted as 0	Predicted as 1
Labeled as 0	164	37
Labeled as 1	58	85

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 15_CatBoost

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.05
• depth: 8
• rsm: 0.8
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

98.9 seconds

Metric details

	score	threshold
logloss	0.617327	nan
auc	0.694256	nan
f1	0.619883	0.373236
accuracy	0.662791	0.426523
precision	1	0.732844
recall	1	0.0847506
mcc	0.313332	0.426523

Confusion matrix (at threshold=0.426523)

	Predicted as 0	Predicted as 1
Labeled as 0	138	63
Labeled as 1	53	90

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 17_CatBoost_RandomFeature

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 7
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

60.7 seconds

Metric details

	score	threshold
logloss	0.640051	nan
auc	0.656786	nan
f1	0.602094	0.360286
accuracy	0.648256	0.507129
precision	0.75	0.604044
recall	1	0.058218
mcc	0.249441	0.507129

Confusion matrix (at threshold=0.507129)

	Predicted as 0	Predicted as 1
Labeled as 0	167	34
Labeled as 1	87	56

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 27_CatBoost_SelectedFeatures

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 8
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

3.2 seconds

Metric details

	score	threshold
logloss	0.60816	nan
auc	0.711443	nan
f1	0.639386	0.332653
accuracy	0.69186	0.532196
precision	0.826087	0.685581
recall	1	0.0216875
mcc	0.349704	0.532196

Confusion matrix (at threshold=0.532196)

	Predicted as 0	Predicted as 1
Labeled as 0	171	30
Labeled as 1	76	67

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 17_CatBoost

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 7
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

67.3 seconds

Metric details

	score	threshold
logloss	0.606855	nan
auc	0.710886	nan
f1	0.628205	0.386881
accuracy	0.68314	0.413698
precision	0.894737	0.704909
recall	1	0.0205809
mcc	0.3484	0.413698

Confusion matrix (at threshold=0.413698)

	Predicted as 0	Predicted as 1
Labeled as 0	146	55
Labeled as 1	54	89

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 23_NeuralNetwork

<< Go back

Neural Network

• n_jobs: -1
• dense_1_size: 32
• dense_2_size: 4
• learning_rate: 0.05
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

8.4 seconds

Metric details

	score	threshold
logloss	2.28946	nan
auc	0.628223	nan
f1	0.604966	5.14916e-06
accuracy	0.630814	0.47208
precision	0.572816	0.714747
recall	1	1.0493e-29
mcc	0.23926	0.47208

Confusion matrix (at threshold=0.47208)

	Predicted as 0	Predicted as 1
Labeled as 0	138	63
Labeled as 1	64	79

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 34_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 0.8
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

7.2 seconds

Metric details

	score	threshold
logloss	0.630705	nan
auc	0.668806	nan
f1	0.619266	0.229776
accuracy	0.659884	0.55527
precision	0.735294	0.660038
recall	1	0.0781465
mcc	0.278825	0.55527

Confusion matrix (at threshold=0.55527)

	Predicted as 0	Predicted as 1
Labeled as 0	182	19
Labeled as 1	98	45

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 6_Default_RandomForest

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.9
• min_samples_split: 30
• max_depth: 4
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

42.1 seconds

Metric details

	score	threshold
logloss	0.6278	nan
auc	0.683401	nan
f1	0.6313	0.358631
accuracy	0.665698	0.496677
precision	0.83871	0.59254
recall	1	0.126216
mcc	0.298001	0.574319

Confusion matrix (at threshold=0.496677)

	Predicted as 0	Predicted as 1
Labeled as 0	172	29
Labeled as 1	86	57

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 8_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.1
• max_depth: 8
• min_child_weight: 1
• subsample: 1.0
• colsample_bytree: 1.0
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

408.7 seconds

Metric details

	score	threshold
logloss	0.630069	nan
auc	0.678339	nan
f1	0.619512	0.311405
accuracy	0.665698	0.493789
precision	0.833333	0.743631
recall	1	0.0749824
mcc	0.298461	0.493789

Confusion matrix (at threshold=0.493789)

	Predicted as 0	Predicted as 1
Labeled as 0	155	46
Labeled as 1	69	74

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 32_LightGBM_SelectedFeatures

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 63
• learning_rate: 0.05
• feature_fraction: 1.0
• bagging_fraction: 0.9
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

3.1 seconds

Metric details

	score	threshold
logloss	0.610673	nan
auc	0.703197	nan
f1	0.640669	0.355324
accuracy	0.677326	0.493598
precision	0.857143	0.727352
recall	1	0.0349144
mcc	0.327961	0.45146

Confusion matrix (at threshold=0.493598)

	Predicted as 0	Predicted as 1
Labeled as 0	168	33
Labeled as 1	78	65

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 10_Xgboost

<< Go back

Extreme Gradient Boosting (Xgboost)

• n_jobs: -1
• objective: binary:logistic
• eta: 0.15
• max_depth: 8
• min_child_weight: 50
• subsample: 0.6
• colsample_bytree: 0.6
• eval_metric: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

408.1 seconds

Metric details

	score	threshold
logloss	0.693147	nan
auc	0.5	nan
f1	0.587269	0.45
accuracy	0.415698	0.45
precision	0.415698	0.45
recall	1	0.45
mcc	0	0.45

Confusion matrix (at threshold=0.45)

	Predicted as 0	Predicted as 1
Labeled as 0	0	201
Labeled as 1	0	143

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 19_RandomForest

<< Go back

Random Forest

• n_jobs: -1
• criterion: gini
• max_features: 0.5
• min_samples_split: 20
• max_depth: 4
• eval_metric_name: logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

26.5 seconds

Metric details

	score	threshold
logloss	0.616654	nan
auc	0.708451	nan
f1	0.628713	0.323098
accuracy	0.697674	0.459418
precision	0.894737	0.66176
recall	1	0.116197
mcc	0.367593	0.459418

Confusion matrix (at threshold=0.459418)

	Predicted as 0	Predicted as 1
Labeled as 0	159	42
Labeled as 1	62	81

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 25_NeuralNetwork

<< Go back

Neural Network

• n_jobs: -1
• dense_1_size: 64
• dense_2_size: 32
• learning_rate: 0.08
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

9.5 seconds

Metric details

	score	threshold
logloss	1.80514	nan
auc	0.617107	nan
f1	0.593607	7.37429e-05
accuracy	0.645349	0.797318
precision	0.684211	0.797318
recall	1	2.39178e-14
mcc	0.242803	0.797318

Confusion matrix (at threshold=0.797318)

	Predicted as 0	Predicted as 1
Labeled as 0	183	18
Labeled as 1	104	39

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 16_CatBoost

<< Go back

CatBoost

• n_jobs: -1
• learning_rate: 0.1
• depth: 8
• rsm: 1.0
• loss_function: Logloss
• eval_metric: Logloss
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

98.6 seconds

Metric details

	score	threshold
logloss	0.629594	nan
auc	0.684445	nan
f1	0.612022	0.309702
accuracy	0.665698	0.493268
precision	0.85	0.67119
recall	1	0.038292
mcc	0.293883	0.421099

Confusion matrix (at threshold=0.493268)

	Predicted as 0	Predicted as 1
Labeled as 0	166	35
Labeled as 1	80	63

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back

Summary of 14_LightGBM

<< Go back

LightGBM

• n_jobs: -1
• objective: binary
• num_leaves: 95
• learning_rate: 0.05
• feature_fraction: 1.0
• bagging_fraction: 1.0
• min_data_in_leaf: 10
• metric: binary_logloss
• custom_eval_metric_name: None
• explain_level: 1

Validation

• validation_type: kfold
• k_folds: 5
• shuffle: True
• stratify: True

Optimized metric

logloss

Training time

6.5 seconds

Metric details

	score	threshold
logloss	0.631686	nan
auc	0.659813	nan
f1	0.617169	0.24762
accuracy	0.642442	0.541502
precision	0.8	0.693257
recall	1	0.0456055
mcc	0.234163	0.345597

Confusion matrix (at threshold=0.541502)

	Predicted as 0	Predicted as 1
Labeled as 0	181	20
Labeled as 1	103	40

Learning curves

Permutation-based Importance

Confusion Matrix

Normalized Confusion Matrix

ROC Curve

Kolmogorov-Smirnov Statistic

Precision-Recall Curve

Calibration Curve

Cumulative Gains Curve

Lift Curve

<< Go back